Problem Set 17 Solutions

Embed Size (px)

Citation preview

  • 7/27/2019 Problem Set 17 Solutions

    1/2

    PROBLEM SET 17 SOLUTIONS.

    (1) Consider the matrixA =

    1 0

    1 2

    .

    (a) Find the eigenvalues ofA.(b) Find the eigenvectors ofA.(c) Diagonalize A: write it as A = PDP1.

    ANSWER:

    (a) We just have to solve the characteristic equation

    det

    1 01 2

    = (1 )(2 ) = 0.

    So = 1, 2. Note that in general the eigenvalues of an upper triangular or lowertriangular matrix are the diagonal entries.

    (b) To find the eigenvector corresponding to , we have to find the null space ofA I. There are two cases.

    = 1. Here we have to solve0 0

    1 1

    a

    b

    =

    00

    So the eigenvector is 11

    = 2. Here we have to solve1 01 0

    a

    b

    =

    00

    So the eigenvector is 01

    (2) Consider the matrix

    A =

    3 4 43 5 3

    1 2 0

    .

    (a) Find the eigenvalues ofA. Just kidding! The eigenvalues are 1, 1 and 2. Twoeigenvectors are

    v =

    21

    0

    & w =

    101

    .

    Check that these are eigenvectors. What are the corresponding eigenvalues?1

  • 7/27/2019 Problem Set 17 Solutions

    2/2

    2 PROBLEM SET 17 SOLUTIONS

    ANSWER: 3 4 43 5 31 2 0

    21

    0

    =

    21

    0

    So the eigenvalue corresponding to v is 1. 3 4 43 5 31 2 0

    101

    =

    101

    So the eigenvalue corresponding to w is 1.(b) Find a third eigenvector corresponding to the third eigenvalue.

    ANSWER: We need to find an element of the null space ofA 2I, in other wordsa solution to

    5 4 43 3 31 2 2

    a

    b

    c

    =

    000

    We can solve this using row reduction, or just observe that the third column is just 1times the second column, and so from the column definition of matrix multiplicationwe know immediately that the eigenvector must be,

    011

    (Note: if you know about vector calculus, another neat way to find the eigenvectoris to take the cross-product of any two rows ofA I.)

    (3) Suppose that A is an nn matrix, and that A2 = A. What can you say, then, aboutthe eigenvalues ofA?

    ANSWER: If is an eigenvalue ofA. Then

    Av = v

    Where v is an eigenvector corresponding to . Therefore

    A2v = AAv = Av = Av = 2v

    And so A2 = A implies A2v = Av . So it must be that

    = 2

    And must be 0 or 1.(4) Suppose A is a 3 3 matrix with eigenvalues 1, 2 and 3. Ifv1 is an eigenvector for

    the eigenvalue 1,v2

    for 2, andv3

    for 3, then what isA

    (v1

    +v2 v3

    )?ANSWER:

    A(v1 +v2

    v3) = Av1 + A

    v2 Av3 = 1

    v1 + 2v2 3

    v3